Measuring Performance: Methods and Error Analysis in Psychology

Measuring Performance in Psychology

Introduction

Understanding how to measure performance is essential in psychological science, especially in the study of learning and behavior. This topic covers the distinction between performance and learning, methods for assessing performance, and quantitative measures of error in behavioral tasks.

Performance vs. Learning

Definitions and Distinctions

• Performance: An observable behavior that can be measured directly (e.g., hitting a ball).

• Learning: A relatively permanent change in performance resulting from practice or experience. Learning is inferred from improvements in performance over time.

• Key Point: Performance is what is seen and measured; learning is the underlying change that leads to improved performance.

• Example: A person may perform well in a single instance (performance), but only repeated improvement over time indicates learning.

Methods for Measuring Performance

Performance Production Measures

Production measures focus on the behavior required to achieve a goal.

• Production: Refers to the execution of movement or behavior (e.g., drawing a bow in archery).

• Assessment Methods:

  • Human Judgement: Observers rate or score performance. Useful for simple tasks but limited for complex or rapid movements.

  • Video Recording: Film or digital recordings allow for detailed analysis and slow-motion review, reducing memory load and improving accuracy.

  • High Speed Filming & Computer Analysis: Enables calculation of position, velocity, and acceleration of body parts by slowing down movement for precise measurement.

  • Electromyogram (EMG): Measures electrical activity in muscles, providing data on the size and timing of muscle contractions during movement.

  • Brain Activity Measures:

    • EEG (Electroencephalogram): Records electrical activity ('brain waves') using electrodes on the scalp.

    • fMRI (Functional Magnetic Resonance Imaging): Measures blood flow to active brain areas, indicating neural activity during performance.

• Example: In archery, drawing the bow is production; accuracy of the shot is outcome.

Performance Outcome Measures

Outcome measures focus on the result of the behavior in relation to a goal or criterion.

• Response Magnitude: The absolute size of the response (e.g., number of laps run, speed of a puck, hours spent learning).

• Reaction Time (RT): Time from stimulus appearance to movement initiation.

• Fractionated Reaction Time (FRT): Breaks RT into components:

  • Premotor Time (PRMOT): Time from stimulus to first EMG activity.

  • Motor Time (MOT): Time from first EMG activity to movement onset.

• Accuracy/Error: Performance relative to a criterion (e.g., shots in basketball, distance from bullseye in darts, timing in video games).

• Example: In golf putting, the distance from the ball to the hole is an outcome measure.

Measures of Error in Performance

Overview

Error measures quantify how closely performance matches a target or criterion. Four main types are used:

• Constant Error (CE)

• Variable Error (VE)

• Total Variability (E or RMS)

• Absolute Error (AE)

Constant Error (CE)

CE measures the average deviation from the target, considering both direction and magnitude.

• Formula: $CE = \frac{\sum (X_i - T)}{n}$ Where $X_i$ is the score, $T$ is the target, and $n$ is the number of trials.

• Example: Scores: 10, 4, -1, -5; Target: 0 $CE = \frac{10 + 4 - 1 - 5}{4} = \frac{8}{4} = 2$ cm

• Interpretation: Shows average distance and direction of errors.

Variable Error (VE)

VE is the standard deviation of scores, indicating variability or inconsistency around the mean performance.

• Formula: $VE = \sqrt{\frac{\sum (X_i - \text{mean})^2}{n}}$

• Example: Mean = 2 cm; VE = 5.6 cm

• Interpretation: Reflects how consistent or variable the performance is relative to one's own average.

Total Variability (E or RMS)

Total variability, or root mean square error (RMS), measures overall error relative to the target.

• Formula: $E = RMS = \sqrt{\frac{\sum (X_i - T)^2}{n}}$

• Example: For scores 10, 4, -1, -5 and target 0: $E = RMS = \sqrt{\frac{100 + 16 + 1 + 25}{4}} = \sqrt{\frac{142}{4}} = 5.96$ cm

• Interpretation: Indicates the spread of scores about the target, serving as an overall measure of success.

Absolute Error (AE)

AE measures overall accuracy, ignoring the direction of errors.

• Formula: $AE = \frac{\sum |X_i - T|}{n}$

• Interpretation: Represents the mean absolute deviation from the target, providing a measure of accuracy without considering direction.

Summary Table: Error Measures

Measure	Formula	What it Represents
Constant Error (CE)	$\frac{\sum (X_i - T)}{n}$	Average distance and direction from target
Variable Error (VE)	$\sqrt{\frac{\sum (X_i - \text{mean})^2}{n}}$	Variability/inconsistency around mean
Total Variability (E or RMS)	$\sqrt{\frac{\sum (X_i - T)^2}{n}}$	Overall error relative to target
Absolute Error (AE)	$\frac{\sum |X_i - T|}{n}$	Mean absolute deviation from target

Summary: Measures of Error

• CE large, VE large: Inaccurate and inconsistent (high variability).

• CE small, VE small: Accurate and consistent (low variability).

• CE small, VE large: Accurate but inconsistent (high variability for direction).

• CE large, VE small: Inaccurate but consistent (low variability).

These combinations help interpret the quality of performance in psychological experiments and applied settings.

Applications in Psychology

Relevance to Learning and Motor Control

• These measures are widely used in experimental psychology, cognitive neuroscience, and motor learning research.

• They help quantify improvement, consistency, and accuracy in tasks ranging from sports to clinical rehabilitation.

• Understanding error measures is essential for designing experiments and interpreting behavioral data.

Next Steps

The next topic will cover the Somatosensory System, which is crucial for understanding sensory feedback in performance and learning.